RU2665157C2 - System and method for assessing risk associated with glycemic state - Google Patents

Abstract

FIELD: medicine.SUBSTANCE: group of inventions relates to medicine, particularly to endocrinology. For continuous monitoring of glycemia by means of a computing device, a target glycemic state is determined, including a target glycemia level and a target rate of change of the target glycemia level; determining the initial glycemic state including an initial glycemia level and an initial rate of change of the initial glycemia level, the initial glycemia state being different from the target glycemia state. By hazard analysis logic of the computing device, a return path for a transition from the initial glycemic state to the target glycemic state is calculated, the return path comprising at least one intermediate glycemic state associated with the transition from the initial glycemic state to the target glycemic state. Target return path is calculated by the hazard analysis logic based on a hypo- or hyperglycemia hazard associated with the at least one intermediate glycemic state of the target return path. By means of the computing device, human treatment is adjusted to transfer a person from the registered glycemic state to a target glycemic state based on the return path and / or hypo- or hyperglycemia hazard metric. At least one hazard metric associated with the registered glycemic state includes the total estimated time of the person's transition from the registered glycemic state to the target glycemic state via the return path, and the total estimated time is calculated based on the specified maximum acceleration of glycemia and the number of intermediate glycemic states on the return path.EFFECT: group of inventions makes it possible to obtain hazard metrics associated with the results of measurements with continuous monitoring of glycemia taking into account the specified maximum acceleration of glycemia, which is adjusted based on the physiology of the patient and the number of intermediate glycemic states on the return path.28 cl, 19 dwg

Description

FIELD OF THE INVENTION

The present invention relates generally to continuous monitoring of glycemia (LMWH) and, in particular, to systems and methods for assessing the risk associated with a glycemic condition.

State of the art

Biological monitoring provides health professionals and patients with biological data that can be used to treat and / or manage a medical condition associated with biological data. For example, devices for continuous monitoring of glycemia (LMWH), also known as continuous glycemic testing, provide data related to the recorded level or concentration of glucose in the blood of people with diabetes (glycemia). In order to assess the risk for a person with diabetes based on the measured glycemia level, risk metrics can be derived from glycemic data. However, existing risk metrics often do not take into account the rate of change of glycemic data and the uncertainty associated with the accuracy of glycemic data. For this reason, existing risk metrics are often unsuitable for use as metrics to optimize treatment or to estimate the total risk for the entire window of measurement results for LMWH.

For example, a known risk metric contains a risk function illustrated in FIG. 1 by schedule 10 and proposed in the following work: Kovatchev, V.P. et al, Symmetrization of the Blood Glucose Measurement Scale and its Applications, Diabetes Care, 1997, 20, 1655-1658 (BP Kovachev et al. "Symmetrization of the glycemic measurement scale and its applications"). The risk function proposed by Kovachev shown in FIG. 1, defined by the equation h (g) = [1.509 (log (g) ^1.0804 -5.381)] ² , where g is the concentration of glucose in the blood (expressed in milligrams per deciliter, or mg / dl), plotted along the abscissa, ah (g) is the corresponding penalty value, pending along the ordinate axis. The value of the penalty (i.e. risk) obtained using the Kovachev function is static in the sense that the penalty depends only on the level of glycemia. Minimal (zero) risk occurs at a value of 112.5 mg / dl corresponding to section 12 of the curve shown in FIG. 1. As the level of glycemia approaches hypoglycemia (section 14), the risk increases much faster than when the level of glycemia approaches hyperglycemia (section 16).

The risk function proposed by Kovachev does not take into account the rate of change in the level of glycemia, as well as the uncertainty associated with the measured glucose level. For example, the risk for a patient associated with a glucose level of 100 mg / dl and its rapid drop will be clearly higher than the risk associated with a glucose level of 100 mg / dl at a constant rate of change. Further, the glycemic measurement results obtained from the glucose sensor may include sensor noise, for example, noise due to physical movement of the glucose sensor relative to the human body, or electrical noise inherent in the glucose sensor. In addition, the glucose sensor may be malfunctioning or may malfunction, for example, due to a failure of the electronics or the battery or due to separation or loss of the sensor. For this reason, the measured glycemia level may not be accurate. The penalty values obtained using the Kovachev function do not take into account such uncertainty in the measured glycemia level.

Accordingly, some embodiments of the present invention provide risk metrics associated with the measurement results for NMH and taking into account the level of glycemia, the rate of change in the level of glycemia and / or the uncertainty associated with the level of glycemia and its rate of change. Further, some embodiments of the present invention provide for calculating a target return path from a given glycemic state to a target glycemic state based on one or more risk metrics associated with intermediate glycemic conditions of the target return path.

Disclosure of invention

As one object of the present invention, a method for analyzing a glycemic state is provided. When implementing the method of the invention, the target glycemic state (target glycemic state) is determined using at least one computing device, including the target glycemic level and the target rate of change of the target glycemic level. In addition, when implementing the method of the invention, at least one computing device determines the initial glycemic state (initial state of glycemia), which includes the initial level of glycemia and the initial rate of change of the initial level of glycemia, whereby glycemia is understood as blood glucose or other physiological fluid. The initial glycemic state is different from the target glycemic state. In addition, when implementing the method of the invention by means of the risk analysis logic available in at least one computing device, I calculate! the target return path for the transition from the initial glycemic state to the target glycemic state. The target return path contains at least one intermediate glycemic state associated with the transition from the initial glycemic state to the target glycemic state. The target return path is calculated by the risk analysis logic based on the risk associated with at least one intermediate glycemic state on the target return path.

As another object of the present invention, a method for analyzing the glycemic state of a person suffering from diabetes is provided. When implementing this method, at least one computing device records the glycemic state of a person based on at least one measured value of glycemia obtained using a glucose sensor. The registered glycemic state includes the level of human glycemia and the rate of change in the level of glycemia. Further, by means of the risk analysis logic available in the at least one computing device, a target return path for transitioning from a registered glycemic state to a target glycemic state is determined. The target glycemic state includes the target glycemic level and the target rate of change of the target glycemic level. The target return path contains at least one intermediate glycemic state associated with the transition from the registered glycemic state to the target glycemic state. Also, through the risk analysis logic available in the at least one computing device, at least one risk metric associated with the registered glycemic state is calculated based on the at least one intermediate glycemic state on the target return path.

As another object of the present invention, there is provided a non-volatile computer-readable storage medium. The non-volatile computer-readable storage medium contains executable instructions, when executed by at least one processor, the specified processor determines the target glycemic state, which includes the target glycemic level and the target rate of change of the target glycemic level. Further, when performing these commands, at least one processor determines the initial glycemic state, which includes the initial level of glycemia and the initial rate of change of the initial level of glycemia. The initial glycemic state is different from the target glycemic state. When these instructions are executed, at least one processor also calculates a return target path for transitioning from the initial glycemic state to the target glycemic state. The target return path contains at least one intermediate glycemic state associated with the transition from the initial glycemic state to the target glycemic state. The target return path is calculated by at least one processor based on the risk associated with the at least one intermediate glycemic state on the target return path.

As another object of the present invention, there is provided a non-volatile computer-readable storage medium. The non-volatile computer-readable storage medium contains executable instructions, when executed by at least one processor, said processor detects a person’s glycemic state based on at least one measured glycemia value obtained using a glucose sensor. The registered glycemic state includes the level of human glycemia and the rate of change in the level of glycemia. Further, when these instructions are executed, at least one processor determines a return target path for transitioning from a registered glycemic state to a target glycemic state. The target glycemic state includes the target glycemic level and the target rate of change of the target glycemic level. The target return path contains at least one intermediate glycemic state associated with the transition from the registered glycemic state to the target glycemic state. When these instructions are executed, at least one processor also calculates at least one risk metric associated with the registered glycemic state based on at least one intermediate glycemic state on the target return path.

Brief Description of the Drawings

Features and advantages of the present invention are clearly presented for specialists in the following detailed description, illustrated by the accompanying drawings, which show:

in FIG. 1 is a known risk function for assessing the risk associated with the level of glycemia;

in FIG. 1A is another exemplary risk function for assessing risk associated with glycemia;

in FIG. 2 is a diagram of a continuous glycemic monitoring system (NMH) used in one or more embodiments of the invention described herein;

in FIG. 3 is a diagram of an exemplary computing device included in FIG. 2 NMG systems and containing the logic of risk analysis;

in FIG. 4 and 5 is a flowchart of an exemplary method for controlling the operation shown in FIG. 3 computing devices for calculating a return path to a target glycemic state from a plurality of glycemic states based on at least one risk metric;

in FIG. 6 is an exemplary penalty matrix that is populated in the implementation of FIG. 4 and 5 of the method and which can function as a lookup table for a given glycemic condition;

in FIG. 7 is a surface graph illustrating, by way of example, the accumulated penalty values for a set of glycemic states calculated by the method shown in FIG. 4 and 5;

in FIG. 8 is a surface graph illustrating, by way of example, the general return times to a target glycemic state from a set of glycemic states calculated by the method shown in FIG. 4 and 5;

in FIG. 9 is a surface graph illustrating, by way of example, the maximum penalty values for a set of glycemic states calculated by the method shown in FIG. 4 and 5;

in FIG. 10 is a surface graph illustrating, by way of example, the average penalty factors for a set of glycemic states calculated by the method shown in FIG. 4 and 5;

in FIG. 11 is a surface graph illustrating, by way of example, the maximum penalty values for the set of glycemic states, signed with the sign, calculated by the method shown in FIG. 4 and 5;

in FIG. 12 is a surface graph illustrating, by way of example, the accumulated penalty values for a set of glycemic states and the probability distribution associated with the glycemic state;

in FIG. 13 is an example graph of the history of NMH results, characterized by low accumulated penalty values for glycemic states included in the history of NMH results;

in FIG. 14 is another example of a graph of the history of NMH results, characterized by moderate accumulated penalty values for glycemic states that are included in the history of NMH results;

in FIG. 15 is another example of a graph of the history of NMH results, characterized by high accumulated penalty values for glycemic states that are included in the history of NMH results;

in FIG. 16 is a flowchart of another exemplary method for controlling the operation of FIG. 3 computing devices for calculating the target return path from the initial glycemic state to the target glycemic state;

in FIG. 17 is a flowchart of another exemplary method for controlling the operation of FIG. 3 computing devices for determining a risk metric associated with a registered glycemic condition; and

in FIG. 18 are three exemplary surface plots showing exemplary accumulated penalty values for a set of glycemic states calculated as shown in FIG. 4 and 5 in a manner based on the risk function shown in FIG. 1A.

The implementation of the invention

To facilitate understanding of the essence of the present invention, variants of its implementation are discussed below, illustrated by the drawings and described using special terminology. It should be borne in mind that the options do not limit the possibilities of carrying out the invention.

The term “logic” or “control logic” in the context of this description may include modifiable and / or embedded (wired) software running in one or more programmable processors, in specialized (custom) integrated circuits (ASICs), in user-programmable ones gate arrays (ППВМ), in digital signal processors (DSP), in hard logic (logic circuits with hard or constant connections) or in combinations of the above. Thus, in accordance with embodiments of the invention, the various logic may be implemented in any suitable manner, while remaining consistent with the embodiments disclosed herein.

In the context of this description, “measured glycemia values” or “glycemia results” means human glycemia levels measured by a glucose sensor, “actual glycemia level” refers to the actual human glycemia level, and “estimated glycemia level” refers to human glycemia level assessment , which can be based on measured glycemia values.

In FIG. 1A shows another exemplary risk function 30 used to calculate the static penalty values for a given glycemia level. The risk function 30 is defined by the following equation:

where g is the glycemia level (mg / dl), laid off along the abscissa axis, h (g) is the corresponding static value of the penalty, laid off along the ordinate axis, ag ₁ and g ₂ are glycemia levels defining the boundaries of the range of target glycemia values (g ₁ ≤ g≤g ₂ ) or one target glycemic value (g ₁ = g ₂ ). In the present embodiment, the variables α, β, and c are defined as follows: α = 1,509, β = 5,381, and c = 1,084. For illustrative purposes, the range of target glycemic values (g ₁ ≤g≤g ₂ ) has a corresponding penalty value of zero, as shown in equation (1). At the target glycemic level g ₁ = g ₂ = 112.5 mg / dl, the risk function 30 gives the risk curve 32 shown in FIG. 1A and the corresponding Kovachev function. At g ₁ = 75 mg / dl and g ₂ = 125 mg / dl, the risk function 30 gives the risk curve 34 shown in FIG. 1A. Moreover, the risk curve 34 allows you to get the fine values for a given glycemic state, when the target range of glycemic values is defined as component from 75 to 125 mg / dl. Other suitable target glycemic levels / ranges and penalty values corresponding to these target glycemic levels / ranges may be provided.

In FIG. 2, as an example embodiment of the invention, a diagram of a continuous glycemic monitoring (NMH) system 50 is provided for monitoring, or tracking, the glycemic level of a person suffering from diabetes. In particular, the NMG system 50 is configured to collect glycemia values measured at a predetermined adjustable interval, for example, every minute, every five minutes, or at other suitable intervals. The NMH system 50, by way of illustration, comprises a glucose sensor 56 having a needle or probe 58 inserted under the skin of a person 52. The tip of the needle 58 is then located in the interstitial fluid 54, such as blood or other physiological fluid, as a result of which the measurements performed by the glucose sensor 56 are based on the glucose level in the interstitial fluid 54. The glucose sensor 56 is installed in the abdomen of a person or in another suitable place. In one embodiment, the glucose sensor 56 is periodically calibrated to increase its accuracy. This periodic calibration may contribute to the correction of the drift of the sensor readings due to the deterioration of the sensor characteristics and changes in the physical condition of the sensor injection site. The glucose sensor 56 may also contain other components, including (without limitation) the wireless transmitter 60 and the antenna 62. Although a glucose sensor 58 is used as an illustration in the glucose sensor 56 to provide access to blood or other human fluid, the glucose sensor 56 can use other suitable instruments for measuring, such as a non-invasive instrument (for example, an infrared light sensor).

When measuring, the glucose sensor 56 transmits the measured glycemia value, i.e. the result of a corresponding measurement, via communication line 64 to a computing device 66, which is, by way of illustration, a glucose monitor 66. Communication line 64 is, by way of illustration, wireless, for example radio frequency or using other frequencies suitable for wireless transmission, and the glycemic measurement results are transmitted over such a communication line via electromagnetic waves. An example of a wireless radio communication system is a Bluetooth® system using a frequency of approximately 2.4 gigahertz (GHz). As another example of a wireless communication scheme, wireless communication systems using infrared light can be mentioned, for example, systems supported by the InfraRed Data Association®, abbreviated IrDA®. Other suitable types of wireless communications may be provided. Communication line 64 may be unidirectional or one-way (i.e., data is transmitted only from glucose sensor 56 to computing device 66) or bi-directional or two-way (i.e., data is transmitted between glucose sensor 56 and computing device 66 in both directions) . In addition, communication line 64 may provide data transmission between two or more devices, for example, between a glucose sensor 56, a computing device 66, a therapeutic device (eg, an insulin pump), and other suitable devices or systems. Although in FIG. 2 shows a wireless communication line 64, as an alternative to it, a wired communication line, for example, a wired communication channel using Ethernet technology, can be provided. Other suitable public or private wired or wireless communication lines may be used.

In FIG. 3 shows an exemplary computing device 66, which is part of the one shown in FIG. 2 systems 50 NMG. Computing device 66 comprises at least one processor 72 executing program code of the variable and / or firmware stored in the memory 76 of computing device 66. The code of the variable / firmware contains instructions by which the processor 72 of computing device 66 is provided the execution of the computing device 66 functions described in the present description. Alternatively, computing device 66 may include one / one or more specialized (custom) integrated circuits (ASICs), user programmable gate arrays (DPCs), digital signal processors (DSPs), hard logic, or combinations of the above. Although, by way of illustration, computing device 66 is a glycemic monitor 66, other suitable computing devices 66 may be provided, such as, for example, desktop computers, laptop computers (laptops), computer servers, personal digital assistants (PDAs), smartphones, cellular devices, tablet computers, infusion pumps (pumps), an integrated device that includes a processor for measuring glycemia and a personal digital assistant or cell phone, etc. Although computing device 66 is presented for clarity as a single computing device 66, several computing devices may be provided that are used together to perform the functions of computing device 66 described herein.

The storage device 76 is any suitable computer-readable storage medium available for the processor 72. The storage device 76 may be a single drive or multiple drives, may be located inside or outside of the computing device 66, and may include both volatile and non-volatile storage media . In addition, the storage device 76 may include removable and / or non-removable storage media. An exemplary storage device 76 includes a random access memory (RAM), read-only memory (ROM), electrically erasable programmable ROM (EEPROM), flash memory, read-only CD (CD-ROM) ), a universal digital disk (DVD), or other optical disk drive, a magnetic disk, or any other suitable medium that is capable of storing data and is accessible to computing device 66.

Computing device 66 also includes an input device 74 electrically coupled to the processor 72. The input device 74 includes any suitable wireless and / or wired communication module configured to transmit data between the processor 72 and the glucose content sensor 56 via communication line 64. In one embodiment, the input device 74 comprises an antenna 70 (FIG. 2) for wirelessly receiving and / or wirelessly transmitting data over a communication link 64. In the present embodiment, the input device 74 is configured to receive data, such as glycemia measurements, from that shown in FIG. 2 of the glucose sensor 56 and providing received data to the processor 72. The computing device 66 stores in the memory 76 the glycemia measurement results received from the glucose sensor 56 through the input device 74.

The computing device 66 also includes a display 68, electrically connected to the processor 72. The display 68 can be built on the basis of any suitable information display technology (for example, based on liquid crystal technology, etc.) with the ability to display information to the user issued by the processor 72 The processor 72 is configured to transmit to the display 68 information related to the registered or estimated glycemic state of a person. The displayed information may include the estimated glycemic state of a person and / or the glycemic state of a person predicted at some point in time in the future. The glycemic state may include an estimated level of glycemia and / or an estimated rate of change in the level of glycemia. The information displayed may also include an assessment of the quality or uncertainty of the estimated glycemia level. In addition, the information displayed may include warnings, alerts, etc., regarding whether the estimated or predicted level of human glycemia is hypoglycemic or hyperglycemic. For example, a warning may be issued if a person’s glycemic level has fallen (or is predicted to fall) below a predetermined hypoglycemic threshold, such as 50 milligrams of glucose per deciliter of blood (mg / dl). Computing device 66 may also be configured to transmit information or warnings to a person in a tactile manner, for example, by vibration.

In one embodiment, the computing device 66 is connected to a remote computing device, such as a remote computing device, at the disposal of a caregiver, or in a place accessible to the caregiver, and data is transmitted between these devices (e.g., data glycemia or other physiological data). In this embodiment, the computing device 66 and the remote device are capable of transmitting physiological data via an information connection, for example, via the Internet or cellular communications, or by physically moving a storage device, such as a floppy disk, a USB storage device (data medium with a universal serial interface bus), a CD or other portable storage device.

As discussed in more detail herein, the processor 72 of computing device 66 comprises risk analysis logic 80 configured to calculate a target return path from each of a plurality of glycemic state data to a target glycemic state. The accumulated, or total, penalty values associated with the target return paths are stored in a matrix, which can be used as a look-up table, as discussed in the present description. As an illustration, the target glycemic state is the optimal, or ideal, glycemic state with no risk, although in principle any suitable target glycemic state can be set. Each target return path consists of many glycemic states that are intermediate, i.e. between the given (current) glycemic state and the optimal glycemic state. In this embodiment, each return path is calculated from the condition of minimizing the total estimated, or estimated, risk associated with intermediate glycemic conditions that are on the return path. Based on the calculated return path by the computing device 66, various control strategies, for example, correction of a person’s treatment, can be applied. In addition, the risk analysis control logic 80 calculates a plurality of risk metrics associated with each given glycemic state based on the calculated return path from the given glycemic state. In the present embodiment, the risk analysis control logic 80 is also configured to analyze the glucose results obtained with the glucose sensor 56 to determine the probability of accuracy of the glucose sensor 56. In addition, the computing device 66 comprises a recursive filter 82 configured to evaluate a person’s glycemic state by weighing the results of glycemia measurement by the probability of accuracy of the glucose sensor, i.e. using that probability. Further, the risk analysis logic 80 is configured to calculate the risk associated with the registered risk, i.e. a measured glycemic state, based on the penalty value associated with this registered glycemic state, and based on the uncertainty of the registered glycemic state, as discussed in the present description.

In FIG. 4 and 5, a flowchart 100 of an exemplary iterative method is performed using the risk analysis logic 80 in processor 72 to calculate a return path to the target glycemic state for each of the many glycemic states based on at least one risk metric. In the present embodiment, the logic 80 calculates the target return path for each glycemic state by filling in the penalty matrix, in each cell or block of which its glycemic state is represented. As discussed in the present description, each glycemic state represented by a matrix cell includes a glycemic level and a rate of change in glycemic level. The target return path consists of many intermediate glycemic states, each of which is represented by a matrix cell. The penalty matrix contains the accumulated penalty value determined for each glycemic state based on the total risk that will have to be faced on the target return path from the corresponding glycemic state, as discussed in the present description.

At step 102 shown in FIG. 4, logic 80 first sets a set of glycemic states for the penalty matrix. In particular, the size, boundaries and step size of the penalty matrix are set in such a way as to determine the set of glycemic states evaluated in the implementation of the method. See, for example, the penalty matrix R, exemplified in FIG. 6. In the matrix R shown in FIG. 6, each column represents a BG glycemic level ranging from 1 to 400 mg / dl, with a step size of 0.5 mg / dl. Each row of the R matrix represents the ABG rate of change in glycemia, ranging from -5 to 5 mg / dl / min (milligrams of glucose per deciliter of blood per minute) with a step size of 0.025 mg / dl / min. Moreover, the resulting matrix R has a size of 799 by 401 (a total of 320,399 cells), each cell representing its own (different from the others) glycemic state, i.e. each cell represents its own combination of BG glycemia level and ABG rate of glycemic change. In order to specify a lower number of glycemic states or additional glycemic states, other suitable boundaries and pitch sizes of the matrix R can be provided. For illustrative purposes, the rows and columns of the matrix R are shown in FIG. 6 abbreviated. As discussed below, in the implementation of the method shown in FIG. 4 and 5, each cell of the penalty matrix R is filled with the accumulated penalty value.

In one embodiment of the method shown in FIG. 4 and 5, logic 80 also fills in additional matrices representing additional risk or danger metrics for the set of glycemic conditions defined for matrix R. These additional matrices include the matrix T of estimated return times, the matrix M of the maximum fines, and the matrix P of average penalty factors, and each of these matrices has the same size, boundaries, and step size as the penalty matrix R specified in step 102. In addition, logic 80 fills in the backward pointer matrix B corresponding to the size and boundaries of the matrices Rice. Each cell of the matrix B of the reverse pointers is filled with a reverse pointer pointing to another cell. In this case, the target path of return from a given glycemic state to the target glycemic state is determined based on the inverse pointer associated with each glycemic state (i.e., each cell) on the target return path. Each cell of the matrix T of return times is filled with the total estimated time required to translate the person’s glycemic indicators from the glycemic state of the corresponding cell into the target glycemic state according to the calculated target return path. Each cell of the matrix M of maximum fines is filled with the maximum accumulated penalty value of all glycemic states on the target return path, starting, inclusively, from the glycemic state of the corresponding cell, and to the target glycemic state. Each cell of the matrix P of average penalty factors is filled with the average estimated penalty coefficient associated with the target return path calculated for the corresponding glycemic state. In one embodiment, each matrix R, T, M, P, and B contains data stored in memory 76 of computing device 66. As discussed herein, the values contained in matrices R, T, M, and P serve as risk metrics for recorded (measured) and / or determined glycemic conditions.

In step 104, the logic 80 initializes the penalty matrix R with the target glycemic state GS _T. In the present embodiment, the target glycemic state GS _T is the optimal glycemic state at a glucose concentration of 112.5 mg / dl and its rate of change of 0 mg / dl / min, as determined by the Kovachev function described above. The target glycemic state of GS _T may include another suitable target glycemic state or range of glycemic conditions. Logic 80 initializes the matrix R, setting the penalty value associated with the cell (R _{112.5, 0} ) of the target glycemic state GS _T to zero. In one embodiment, logic 80 also initializes matrices T, M, and P, setting the target glycemic state to the appropriate time value, maximum penalty value, and average penalty factor of zero. In one embodiment, logic 80 also initializes all other glycemic states (cells) of matrix R with a large value, such as 100,000 or another suitable large number.

At step 106, the logic 80 initializes the Q queue defining the cells to be evaluated. At the first iteration of the method, the logic 80 to initialize the Q queue adds the target glycemic state GS _{T to the} Q queue. In this case, after step 106, the Q queue initially determines the only cell to be evaluated, i.e. cell corresponding to the target glycemic state of GS _T. At step 108, logic 80 increments the time counter t, i.e. increases its value by a given time step. In one embodiment of the invention, the contents of the time counter t are initially zero, and in step 108, the logic 80 increments the time counter by one minute. In one embodiment of the invention, the time step is set to a small value (for example, for one minute), as a result of which the discrete analysis steps performed during the implementation of the method approach a continuous system. Other suitable time increments may be implemented. At step 110, the logic 80 clears the temporary queue Q _TEMP, used to store these states glycemic GS _N, then added into the queue Q to assess the process of the invention, as discussed herein.

At step 112, logic 80 selects the glycemic state of GS _Q from the Q queue to evaluate. In the initial iteration, in step 112, the selected glycemic state of GS _Q is the target glycemic state of GS _T. In subsequent iterations, the Q queue contains additional glycemic states available for selection at step 112 in order to carry out the assessment considered in the present description. When determining the glycemic state of interest, GS _Q, logic 80 defines a set of glycemic disturbances that can occur during a time step, which is performed at step 114 of the method. Perturbations are determined based on accepted physiological limitations associated with the glycemic state. A set of perturbations is used to determine other potential glycemic states in which a person may end up during a time step (for example, within one minute) upon transition from the glycemic state of interest GS _Q. In other words, the magnitude of the change in the glycemic state of a person’s blood within one minute (or another suitable time step) is naturally limited by physiological limitations. In this case, the logic 80 defines a set of perturbations based on at least one accepted maximum value of the perturbation that may occur during a time step. Based on this accepted maximum perturbation value, the logic 80 in step 114 determines a set of perturbation values that fall within the range determined by the maximum perturbation value.

In the exemplary embodiment of the invention, the disturbances specified in step 114 are the acceleration values associated with the glycemic level. In this example, the logic 80, based on physiological limitations, takes the maximum acceleration threshold and, based on the accepted maximum acceleration and the glycemic state of interest GS _Q , calculates several other potential glycemic states that may occur during the time step. An example of a maximum acceleration threshold is ± 0.025 mg / dl / min ² . At that, at step 114, logic 80 sets a set of acceleration values so that they are in the range from -0.025 mg / dl / min ² to +0.025 mg / dl / min ² . From a given range, logic 80 selects a plurality of discrete accelerations to use as the aforementioned set of acceleration values. An example of a set of acceleration values is the following set: [-0.025, -0.020, -0.015, -0.010, -0.005, 0.000, +0.005, +0.010, +0.015, +0.020, +0.025] (mg / dl / min ² ).

The maximum acceleration can be adjusted taking into account various metabolic conditions of a person with diabetes. In one embodiment, the maximum acceleration is set so as to substantially correspond to the physiology of the patient. For example, in a child, glycemia levels may fluctuate at a faster rate than in an adult. In this regard, a larger maximum acceleration may be suitable for people with a faster metabolism (for example, for children), and a lower maximum acceleration may be suitable for people with a slower metabolism (for example, for adults). An example of an upper threshold for maximum acceleration is ± 0.025 mg / dl / min ² , and an example of a lower threshold for maximum acceleration is ± 0.020 mg / dl / min ² , although other suitable maximum accelerations may be used.

At step 116, the logic 80 determines or selects for estimation from a given set of perturbation values a certain perturbation value (for example, an acceleration value). Based on the glycemic state of GS _{Q of} interest and the perturbation value determined in step 116, logic 80 in step 118 determines the next glycemic state of GS _N to be estimated by the proposed method. For example, if the indicated disturbance is, for example, acceleration, logic 80 determines the following glycemic state GS _N , which includes the level of glycemia and its corresponding rate of change, based on the known level of glycemia and the known rate of change of the selected glycemic state GS _Q , as well based on the acceleration values selected in step 116. for example, logic 80 computes the glucose level G _N and the change rate dG _N glucose included in the following glycemic status GS _N, using the following Eyes:

where G _Q is the glycemic level included in the glycemic state of GS _Q , dG _Q is the rate of change of the glycemic state of GS _Q , dt is the time step set in step 108 of the method (for example, one minute), and a is the acceleration value set in step 116. In one embodiment, the logic 80 rounds the calculated values of G _N and dG _N to the nearest step size determined by the cells of the matrix R. For example, the glycemia values in the cells of the matrix R, shown as an illustration in FIG. 6 have a step size of 0.5 mg / dl, and the rate of change in glycemia has a step size of 0.025 mg / dl / min. At step 120, logic 80 determines whether the next glycemic state GS _N , defined by the rounded values of G _N and dG _N , fits into the boundaries of the matrix R, i.e. whether the cell of the matrix R corresponds to the following glycemic state GS _N.

If it is determined in step 120 that the next glycemic state of GS _{N is} not in the matrix R, logic 80 skips step 122 and proceeds to step 124. If, in step 120, it is established that the next glycemic state of GS _N is in the matrix R, logic 80 proceeds to step 122 to evaluate the risk associated with this next glycemic state of GS _N. At step 122, logic 80 determines whether the accumulated penalty value associated with the next glycemic state of GS _N exceeds the sum of the accumulated value of the fine of the glycemic state of interest GS _Q and the static value of the penalty of the next glycemic state GS _N. In the present embodiment, the static value of the fine of the glycemic state is obtained using the Kovachev function described in the present description. In yet another embodiment, a static glycemic penalty is obtained using the risk function 30 described with reference to FIG. 1A. Other suitable risk functions may be used. The accumulated value of the penalty of a given glycemic state is the sum of the static value of the penalty of this given glycemic state and the static values of the penalty of each intermediate glycemic state determined during the implementation of the method on the target return path associated with this glycemic state. In this case, the accumulated penalty value of the GS _Q glycemic state of interest is the sum of the static value of the GS _Q penalty and the static values of the penalty of all intermediate glycemic states associated with the calculated target return path (during the previous iterations of the method) for this glycemic state GS _Q. For the first iteration of the method, GS _Q is the optimal glycemic state of GS _T , and therefore its accumulated penalty value is zero. For subsequent iterations, the GS _Q can be any other glycemic state of the matrix R having the accumulated penalty value associated with it, which was calculated at previous iterations of the method shown in FIG. 4 and 5. Similarly, the accumulated penalty value of the next glycemic state of GS _{N is} based on a previously calculated target return path associated with this state of GS _N. Thus, if the next glycemic state of GS _N has already been evaluated at any previous iteration of the method, GS _N will be associated with it, i.e. the target return path corresponding to it, and hence the accumulated penalty value associated with it. If, at the current iteration of the method, the next glycemic state of GS _N was first evaluated by the proposed method, then the next glycemic state of GS _{N will} no longer have a target return path associated with it. For the first iteration of the method the accumulated value of the fine state GS _N is the static value of the fine state GS _N.

In this case, at step 122, the logic 80 determines whether the accumulated penalty of the target return path previously calculated for the next glycemic state GS _N exceeds the accumulated penalty of the target return path that is currently being evaluated for GS _N , i.e. cumulative penalty of the return path for GS _Q plus a static risk value for GS _N. If so, then logic 80 determines that a more favorable return return path has been found for GS _N (i.e., a path with a lower accumulated penalty value). In this case, the logic 80 designates this new target return path for GS _N as the currently estimated target return path for GS _Q plus the transition from GS _N to GS _Q. In particular, when the result is “true” in step 122, the method proceeds to that shown in FIG. 5 to step 140, in which the logic 80 sets the back pointer from the GS _N to the GS _Q , thereby linking the GS _N to the destination return path specified for the GS _Q. Logic 80 sets the inverse pointer located in the cell of matrix B corresponding to the next glycemic state GS _N. At step 142 logic 80 sets the accumulated value of the fine state GS _N in the matrix R equal to the sum of the accumulated value of the penalty GS _Q state and the static state value of fine GS _N. In step 144, the logic 80 sets in the matrix T the total estimated return time for GS _N equal to the value of the time counter t. In this case, the total estimated return time set in step 144 is estimated, i.e. estimated time required to return to the target glycemic state GS _T from the next glycemic state GS _N according to the new target return path set for GS _N in step 140. During the method, the time counter t is incremented for each estimated glycemic state on the target return path for GS _N. For example, if the target return path for GS _N set in step 140 contains four intermediate glycemic states between the GS _N and GS _T states, then the time counter t will be equal to five increments, or increments (including the increment from the GS _Q state to the GS _N state ) So, in this example, the total return time calculated in step 144 will be five minutes (given the equal time step of one minute).

At step 146, logic 80 calculates the average penalty factor associated with the target return path for GS _N. The average penalty factor for GS _{N is} calculated by dividing the accumulated penalty value set in step 142 by the total time set in step 144. Logic 80 sets the calculated average penalty coefficient in the matrix cell P to the next glycemic state GS _N. At step 148, logic 80 determines whether the static value of the glycemic penalty GS _N exceeds the current maximum static penalty associated with the destination return path for GS _N. In particular, if the static value of the penalty for the state GS _{N is} greater than the static value of the penalty for each intermediate glycemic state on the target return path for GS _N , then logic 80 in step 150 sets the maximum static value of the penalty associated with the target return path for GS _N equal to the static value fine state GS _n . If the result is “false” in step 122, then logic 80 sets the current maximum penalty value associated with the target by returning from the state GS _Q , for GS _N , i.e. in the cell of the matrix M corresponding to the state GS _N , the logic 80 sets the maximum static value of the penalty.

At step 152, the logic 80 determines whether the next glycemic state GS _N is stored in the Q _TEMP time queue. If not, logic 80 stores the state of GS _N in Q _TEMP and proceeds to that shown in FIG. 4 to step 124 of the method. If the GS _{N is} already stored in the Q _TEMP (i.e., if the GS _N has already been evaluated after clearing the Q _TEMP in step 110), the process proceeds to step 124. In step 124 shown in FIG. 4, logic 80 determines whether the perturbation value (for example, the acceleration value) determined in step 116 was the last value from the set of perturbation values specified in step 114. If additional perturbation values in the set of perturbation values have not yet been evaluated, the method returns to step 116 to select from a specified set a different perturbation value for evaluation with respect to the glycemic state of GS _Q. Then, in the method in step 118, another next glycemic state GS _{N is} calculated based on this new perturbation value, and steps 120-124 are repeated. After all the disturbance values from the set were evaluated in step 124, the logic 80 proceeds to step 126 to determine whether all the glycemic states determined by the Q queue were evaluated. If the additional glycemic states in the Q queue were not evaluated, the method returns to step 112 to select from the queue Q another glycemic state of interest GS _Q for evaluation. Then, logic 80 repeats steps 114-124 for each glycemic state from the Q queue. After all glycemic states in the Q queue are evaluated at step 126, logic 80 clears Q queue and sets Q queue to Q _TEMP in step 128, i.e. e. all glycemic states added to Q _TEMP are placed in the Q _queue (at step 154 of the method shown in Fig. 5). Moreover, if it turns out at step 130 that the Q queue is not empty, the logic 80 returns to step 108 and evaluates all glycemic states placed in the Q queue at step 128. After it turns out at step 130 that the Q queue is empty, which indicates that in step 120, none of the following glycemic states of GS _N evaluated at the last iteration of the method was within the boundaries of the matrix R, or that no new optimal return paths were found for any of the following glycemic states of GS _N , the execution of the method ends.

In one possible embodiment of the invention, the execution of the method is completed after all matrices R, T, M, P and B are completed. If, after completion of the method, one or more cells of the matrix R are left blank, the accumulated penalty value for these corresponding glycemic states can be established equal to the largest penalty contained in the R matrix with the same “+” or “-” sign (ie, the sign corresponding to the risk of hypo- or hyperglycemia). In yet another embodiment, the unfilled glycemic states (cells) of the matrix R can be set to a value that exceeds the highest penalty value. In yet another embodiment, the unfilled glycemic states (cells) of the matrix R can be defined as fault tolerant states leading to a warning.

In one possible embodiment of the invention, the calculated matrices R, T, M and P are used to construct surface or contour graphs (surface graphs with contour lines) that reflect the values of the risk or hazard metric in the corresponding matrices R, T, M, and P. See, for example, surface plots with contour lines shown by way of example in FIG. 7-11, the surface illustrating the corresponding value of a risk or hazard metric. The surface plots with contour lines shown in FIG. 7-11 are constructed using the method shown in FIG. 4 and 5, based on the static penalty values provided by the Kovachev function (see step 122 of the method shown in FIG. 4 and described above). Although for illustrative purposes, the surface graphs shown in FIG. 7-11 are contour plots, you can also build color surface plots on which the color / tone of the surface indicates the corresponding metric value. In FIG. 7, surface 200 of the accumulated fines represents the accumulated penalty values calculated by logic 80 for the glycemic states of the matrix R. The ordinate shows the glycemia level, the values of which are in the range from 0 to 600 mg / dl, and the abscissa axis represents the rate of change in glycemia, the values of which are in the range from -5 to 5 mg / dl / min. Although the R matrix is described above as containing glycemic values in the range of 0 to 400 mg / dl, the surface plots shown in FIG. 7-12, for illustrative purposes, encompass glycemic values in the range of 0 to 600 mg / dl. At point A, an glycemic state is indicated as an example, characterized by a glycemic level of 225 mg / dl and a rate of glycemic change of -1.0 mg / dl / min (see also Figs. 8-10). The return target path 202 is shown leading from the glycemic state at point A to the optimal glycemic state at point O. The return target path 202 calculated from the condition of minimizing the accumulated penalty value according to the method shown in FIG. 4 and 5, illustrates intermediate glycemic states corresponding to the calculated transition from the glycemic state at point A to the optimal glycemic state at point O. Similarly, in FIG. Figure 8-10 shows surface graphs for other risk metrics. In particular, in FIG. 8, the total return time surface 210 represents the total estimated return times calculated by logic 80 for all glycemic states of the matrix T. FIG. 9, surface 220 of maximum fines represents the maximum penalty values calculated by logic 80 for all glycemic states of matrix M. FIG. 10, surface 230 of the penalty factors represents the average penalty factors calculated by logic 80 for all glycemic states of matrix P.

In FIG. 18 shows additional surface graphs with contour lines, each of which corresponds to its own accumulated fines matrix R formed by the method shown in FIG. 4 and 5. Referring to FIG. 18 charts are plotted based on static penalty values obtained using risk function 30 (see step 122 shown in FIG. 4 and described above). Moreover, in the FIGS. 18 graphs depict various surfaces 352, 354, 356 of accumulated fines for various ranges of target glycemic levels (i.e., g ₁ , g ₂ ) considered as examples. The ordinate axis of each surface 352, 354, 356 shows the level of glycemia, the values of which are in the range from 0 to 600 mg / dl, and the abscissa axis represents the rate of change of glycemia, the values of which are in the range from -5 to 5 mg / dl / min In FIG. 18, the surface 352 accumulated fines represents the accumulated penalty values for glycemic states based on the target glycemia value g ₁ = g ₂ = 112.5 mg / dl, and in the region 353 the accumulated penalty is minimal (for example, equal to zero). The surface of 354 accumulated fines represents the accumulated penalty values based on the target range of glycemia values from g ₁ = 80 mg / dl to g ₂ = 120 mg / dl, and in the region 355 the accumulated penalty is minimal (for example, equal to zero). The surface of 356 accumulated fines represents the accumulated penalty values based on the target range of glycemia values from g ₁ = 110 mg / dl to g ₂ = 140 mg / dl, and in the region 357 the accumulated penalty is minimal (for example, equal to zero). Other suitable target ranges of glycemia values may be provided.

Based on the method shown in FIG. 4 and 5, logic 80 is also capable of calculating, for the matrices R, M, P, and B, risk / hazard metrics with a sign. In one embodiment of the invention, to calculate the metrics with a sign, logic 80 sets the static penalty values associated with hypoglycemic conditions, i.e. glycemic conditions with a glycemic level of less than 112.5 mg / dl, negative using the following equation:

where g is the level of glycemia, and H _S (g) is the static value of the penalty with a sign associated with the level of glycemia g. Logic 80 computes the target return path according to the method shown in FIG. 4 and 5, analyzing the absolute value of accumulated fines with a sign. For example, in step 122 shown in FIG. 4, the logic 80 determines whether the absolute value of the accumulated penalty value associated with the GS _N exceeds the absolute value of the sum of the accumulated value of the GS _Q state penalty and the static value of the GS _N state penalty. Similarly, in step 148 of the method shown in FIG. 5, logic 80 determines whether the absolute value of the static value of the state penalty GS _N exceeds the absolute value of the current maximum static value of the penalty associated with the destination return path for GS _N. Based on the values of the penalty with a sign, logic 80 is capable of constructing for each matrix R, M, and P surfaces that represent risks with a sign (positive and negative). For example, in FIG. 11 shows a sign maximum surface 240 of fines allowing a distinction to be made (for example, to illustrate possibilities using color / tone) between negative maximum fines related to the hypoglycemic area and positive maximum penalty values related to the hyperglycemic area.

The computing device 66 shown in FIG. 3 is also configured to estimate a person’s glycemic state based on the results of glycemia measurement obtained with the glucose sensor 56. In particular, the glucose sensor 56 may not function normally due to a malfunction and / or noise associated with the glucose sensor 56, which could potentially cause inaccuracies in the measurement of glycemia. In this case, the risk analysis logic 80 available in the computing device 66 is also capable of analyzing the probability of the accuracy of the recorded glycemic state established by the glucose sensor 56. Logic 80 may use any suitable probabilistic analysis tool to determine the probability of accuracy of a glycemic measurement result, such as a hidden Markov model. Based on a certain accuracy probability, the risk analysis logic 80 estimates the glycemia level and the rate of change in human glycemia using a recursive filter 82 (FIG. 3). In particular, a recursive filter 82, such as a Kalman filter, for example, weighs the recorded glycemic state, including the level of glycemia and its rate of change, using the found probability of accuracy of the glucose sensor. Further, based on the accuracy probability of the glucose sensor, the recursive filter 82 is able to calculate the measure of uncertainty of the estimated glycemic state. A measure of uncertainty is an indicator of the quality of the estimated glycemic state. For a series of recorded glycemic conditions, the uncertainty can vary from one state to another. For a further description of the probabilistic analysis tool, recursive filter, uncertainty calculation, and other functionalities of computing device 66 related to the analysis of probabilities and risks, see US application No. 12/693701 for methods and systems for processing the results of measurements of glycemia in a person with diabetes, the entire contents of which are incorporated into this description by reference.

Depicted in FIG. 12, the surface 250 of accumulated fines represents the accumulated penalty values calculated by logic 80 for the glycemic states of the matrix R, as discussed in the present description. When registering a glycemic state with a glycemic level and a rate of glycemic change corresponding to point B in FIG. 12, logic 80 is capable of calculating a probability distribution around a registered glycemic state. In FIG. 12, two alternative distributions 252 and 254 are shown. A smaller distribution 252 indicates less uncertainty associated with the recorded glycemic state, and a larger distribution 254 indicates greater uncertainty. Distributions 252 and 254, by way of illustration, are Gaussian (normal) distributions, although other suitable methods for representing uncertainty may be provided, such as, for example, a partial filter or a mixture of Gaussians.

Based on the uncertainty associated with the registered glycemic state, the risk analysis logic 80 is able to calculate the risk value for this registered glycemic state. In particular, the risk value is equal to the accumulated penalty of the registered glycemic state provided by the matrix R times the probability of the accuracy of the results of glycemic measurement determined by logic 80. For a given accumulated penalty of the registered glycemic state, the risk value calculated by logic 80 increases with increasing uncertainty of the registered glycemic state. For example, shown in FIG. 12, distribution 252 has a lower risk value than distribution 254, since there is less uncertainty for distribution 252. In the present embodiment, the calculated risk value can be displayed on the display 68 of the computing device 66. In addition, the calculated risk value can be used to correct the treatment provided to a person suffering from diabetes, for example, correction of a bolus and / or basal rate of insulin delivery. For a further description of the functionality of computing device 66 regarding risk calculation and probability distribution calculation, see US Application No. 12/818795 for the invention of "Optimization of Insulin Therapy and Research Methods Using the Corrected Exit Criteria That Considers System Noise, filed June 18, 2010" associated with biomarkers ", the entire contents of which are incorporated into this description by reference.

In FIG. Figures 13-15 show several graphs 260, 270, 280, each of which, as an example, presents the history of the results of continuous glycemic monitoring (NMH), with the time in minutes plotted on the abscissa axis and the glycemia level in mg / dl plotted on the ordinate axis. Each history of LMWH results contains a series of recorded glycemia levels measured over a period of time, and thus illustrates the dynamics of glycemia levels over time. In FIG. 13, graph 260 is shown as an example, showing a curve 261 of unprocessed (unfiltered) glycemic measurement results and a curve 262 of filtered glycemic measurement results (i.e., the glycemic levels on curve 262 are estimated taking into account the probability of sensor accuracy). Each estimated glycemic level on curve 262 has a corresponding fine point 264, the size (diameter) of which represents the accumulated penalty associated with the glycemic level. As shown in FIG. 13, curve 262 passes, basically remaining in the region of 110 mg / dl (near the optimal glycemic level) at a minimum rate of change in glycemic level, and therefore each estimated glycemic level has a low accumulated penalty. Also in FIG. 13 shows a target return path 266 containing intermediate glycemic levels that are passable with a targeted return to an optimal glycemic level of 112.5 mg / dl, and calculated by logic 80 as described herein. Target return path 266 begins with the last estimated glycemic value on curve 262, by way of illustration, in the 240th minute region.

In FIG. 14, graph 270 is shown as an example, showing a curve 271 of the raw glycemic measurement results and a curve 272 of filtered (estimated) glycemic results, each estimated glycemic value on curve 272 having a corresponding fine point 274, the size of which represents the accumulated penalty for the corresponding level glycemia. As shown in the graph, curve 272 grows in the direction of hyperglycemia, but the rate of change in glycemia levels on curve 272 is slow to moderate. Moreover, with an increase in the estimated glycemia levels, the fine points 274 increase in size, but the size of these fine 274 points remains moderate. In addition, the accumulated penalty values are represented by curve 278, which shows that the accumulated penalty reaches a maximum shortly before reaching the peak glycemic level with an increase in glycemia with a positive rate of change. Also in FIG. 14 shows a return target path 276 found for the last estimated glycemia level on curve 272 and calculated by logic 80 as described herein. As shown in the graph, the target return path 276 has a larger estimated return time than the target return path 266 shown in FIG. 13.

In FIG. 15, a graph 280 is shown as an example, showing a curve 282 of the filtered (estimated) glycemic results, each estimated glycemic value on curve 282 having a corresponding fine point 284, the size of which represents the accumulated penalty for the corresponding glycemia level. As shown in the graph, curve 282 goes down in the direction of hypoglycemia, and the rate of change in glycemia levels is higher than the rate of change in curve 272 shown in FIG. 14. At the same time, during the rapid drop in glycemia levels, the fine points 284 increase in size and become relatively large. In addition, the accumulated penalty values are represented by curve 288, which shows that the accumulated penalty reaches a maximum (peak 290) to the lowest glycemia level (point 292), when the rate of change in the glycemia level still drops. At the same time, the maximum accumulated penalty at peak 290 reflects the forecast of future low glycemia levels, and hence the future risk due to a rapid drop in glycemia level (determined by the recorded rate of change of the latter). Also in FIG. 15 shows a return target path 286 found for the last estimated glycemia level on curve 282 and calculated by logic 80 as described herein.

Logic 80 can also calculate the total value of the fine J for the history of the results of the NMH, determined on the basis of the following equation:

where f ₁ is the accumulated penalty of a given glycemic state, which is included in the history of NMH results, i.e. available on the corresponding chart, f ₂ is the accumulated penalty of the last glycemic state included in the history of NMH results, g is the level of glycemia, dg is the rate of change in glycemia, and μ is the parameter used to adjust the balance between the accumulated penalty of the state history and the accumulated penalty of the last state. Moreover, the total penalty J of the NMH results history is the sum of the accumulated fines for all points on the NMH results history graph plus the accumulated penalty for the last state. Alternatively, the functions f ₁ and f ₂ may be other penalty functions described herein, such as the average penalty ratio or the maximum accumulated penalty, or they may be a combination of penalty functions described in the present description.

In FIG. 16 is a flowchart 300 showing an embodiment of an exemplary method performed by FIG. 3 logic 80 risk analysis to calculate the target path return from the initial glycemic state to the target glycemic state. When considering the flowchart of FIG. 16, reference is made to the method of FIG. 4 and 5. At step 302, logic 80 determines the target glycemic state, which includes the target glycemic level and the target rate of change of the target glycemic level. In one embodiment, the target glycemic level is the optimal glycemic level, which is determined by the method shown in FIG. 4 and 5 and described above, and for which the value of the corresponding penalty / risk is zero. At step 304, logic 80 determines the initial glycemic state, which includes the initial level of glycemia and the initial rate of change of the initial level of glycemia. The initial glycemic state is different from the target glycemic state. In one embodiment, the initial glycemic condition is the following glycemic status GS _N, evaluated by the method shown in FIG. 4 and 5. In another embodiment, the initial glycemic state is a glycemic state recorded based on the results of glycemia measurements obtained from the glucose sensor 56 (FIG. 2).

At step 306, the logic 80 calculates a target return path for transitioning from the initial glycemic state to the target glycemic state. As discussed in the present description, the target return path contains at least one intermediate glycemic state associated with the transition from the initial glycemic state to the target glycemic state. The target return path is computed by logic 80 based on the risk associated with at least one intermediate glycemic state on the target return path, as discussed herein. In one embodiment of the invention, logic 80 determines between the initial glycemic state and the target glycemic state a plurality of potential intermediate glycemic states and selects from this set at least one intermediate glycemic state from the condition of minimizing the risk associated with the target return path. For example, to find the target return path with a minimum accumulated penalty, logic 80, as shown in FIG. 4 and 5, several times evaluates the same following glycemic states of GS _N when evaluating various glycemic states of interest of GS _Q during the implementation of the method. Logic 80 then assigns each state GS _N to a target return path having a minimum accumulated penalty, as discussed in the present description.

In one possible embodiment of the invention, the target return path is calculated in step 306 also taking into account the physiological limit of glycemic disturbance, such as a given maximum glycemic acceleration, as discussed in the present description. In one embodiment of the invention, logic 80 in step 306 calculates the target return path by determining, based on the target glycemic state, the set of potential glycemic conditions (GS _N ), the physiological limit of glycemic perturbation (e.g., accepted maximum acceleration), and a predetermined period (e.g., temporary of the step used in step 108 of the method shown in Fig. 4 to increment the time counter), as described herein with respect to FIG. 4 and 5. For example, based on the physiological limit, logic 80 makes the assumption that the transition to the target glycemic state from each of the potential glycemic states is achievable for a person within a given period of time.

In one possible embodiment of the invention, logic 80 calculates a target return path for a plurality of initial glycemic states (e.g., glycemic states contained in matrix R), each target return path being computed by logic 80 from a risk minimization condition (i.e., a penalty value) associated with intermediate glycemic conditions on the target return path, as discussed in the present description. In one embodiment, logic 80 creates one or more lookup tables that store the values of the matrices R, T, M, P, and B for each glycemic state. A lookup table can be used to analyze various risks or dangers associated with a particular glycemic condition of interest. For example, after the glycemic state of a person is recorded by the NMG system 50 (Fig. 2), the matrices R, T, M, P, and B formed during the calculations are stored in the look-up table (which is stored, for example, in the storage device 76 shown in FIG. 3), you can contact to find the risk metric associated with the stored glycemic state, which basically coincides with the registered glycemic state. In one embodiment, the logic 80 is configured to search the following information in a look-up table: the accumulated penalty value associated with the registered glycemic state and obtained from the matrix R, the estimated return time for the registered glycemic state obtained from the matrix T, the maximum penalty, associated with the target return path for the registered glycemic state and obtained from the matrix M, and the average penalty ratio associated with the target put return and obtained from matrix P. Logic 80 also determines the optimal, or target, return path for the registered glycemic state based on the correspondence provided by the reverse pointers of matrix B. In one embodiment of the invention, logic 80 displays a certain risk metric on display 68 shown in FIG. 2, or transfers it to a remote computing system.

In one of the possible embodiments of the invention, the logic 80, based on various maximum glycemic disturbances (specified in step 114 shown in Fig. 4), calculates several sets of matrices R, T, M and P and thus forms several lookup tables, each of which corresponds to to its set of matrices R, T, M, and P. For example, logic 80 calculates its look-up table for each of the many different maximum glycemic accelerations given in FIG. 4 step 114, and, accordingly, each such look-up table contains a unique set of risk or hazard metrics corresponding to the maximum glycemic acceleration associated with the table. Then, each look-up table can be used to analyze risks or dangers. In one embodiment of the invention, the computing device 66 selects from the group of look-up tables a specific look-up table for risk analysis based on at least one user-defined parameter input to the NMH system 50 (Fig. 2) or programmable when it is configured. For example, through the user interface of computing device 66, a user may enter their age or some other suitable parameter. As explained above, the age of a person with diabetes may be important in choosing the right maximum acceleration of glycemia. Based on the input (s) of the parameter (s), logic 80 selects a look-up table corresponding to this parameter (e.g. age) according to the maximum glycemic acceleration associated with the selected look-up table. The selected look-up table can then be used to calculate risk metrics for a person’s registered glycemic state, as discussed in the present description.

In FIG. 17 is a flowchart 310 illustrating the implementation of yet another exemplary method carried out by means of that shown in FIG. 3 logic 80 risk analysis to assess the risk associated with a registered glycemic state. At step 312, logic 80 senses a person’s glycemic state based on at least one measured glycemic value obtained using glucose sensor 56, as discussed herein. At step 314, the logic 80 determines the target return path for the transition from the registered glycemic state to the target glycemic state, as discussed in the present description. In one embodiment, the logic 80 in step 314 determines the target return path by finding the glycemic state closest to the registered glycemic state in the lookup table. The target return path associated with the found nearest glycemic state of the lookup table is then defined as the target return path for the registered glycemic state. At step 316, logic 80 calculates at least one risk metric associated with the registered glycemic state based on at least one intermediate glycemic state on the target return path. In one embodiment of the invention, at least one risk metric is calculated by selecting from a lookup table stored in memory 76 and described above a risk metric associated with a registered glycemic condition. For example, a risk metric may include the accumulated penalty value, the total estimated time required to return from the registered glycemic state to the target glycemic state, the average penalty ratio associated with the target return path, and the maximum penalty value associated with the target return path. In one embodiment, the at least one risk metric is an accumulated risk value calculated by logic 80, as described below.

In particular, the look-up table is also used to account for the uncertainty of the reported glycemic state when analyzing the risk associated with the registered glycemic state. In one embodiment of the invention, logic 80 calculates the risk associated with the registered glycemic state and with all other glycemic conditions contained in the matrix R of the lookup table. Then, logic 80 summarizes all these individual risk values, determining (at step 316) the accumulated risk associated with the registered glycemic state. For example, after registering a person’s glycemic state in step 312, logic 80 calculates the probability of a person being in that registered glycemic state, as described above. In addition, logic 80 calculates the probability of a person being in each of these other glycemic states of the fines matrix R, for example, based on the probability distribution of the registered glycemic state described above. In one embodiment, calculating the probability of each glycemic state includes calculating the probability or uncertainty of the glycemic level and the probability or uncertainty of the rate of change of glycemia for each glycemic state. Then, taking the probability calculation result into account, logic 80 calculates the risk associated with each glycemic state contained in the matrix R, including the registered glycemic state. As described above, each risk value is calculated based on the product of the probability measure and the corresponding accumulated penalty value for the glycemic state. In conclusion, logic 80 summarizes all calculated risks of glycemic conditions contained in matrix R to determine the total, or accumulated, risk associated with the registered glycemic state. The accumulated risk value can be stored in a storage device 76 (Fig. 3), can be presented to the user on the display 68 (Fig. 3) and / or can be used for additional studies or implementation of management strategies.

Alternatively, logic 80 can calculate the probability and the corresponding risk not for all glycemic states contained in the matrix R, but for each of a subset of the glycemic states contained in the matrix R (for example, glycemic states located near a registered glycemic state or within a certain range probability distribution). In addition, the calculation of accumulated risk can be applied to other risk metrics, such as risk metrics provided for in the other penalty matrices discussed in the present description (for example, in the matrix M, P or T).

Based on the target return path determined for the registered glycemic state of a person by a computing device 66, another system, or through human intervention, various control strategies can be applied. For example, computing device 66 may be associated with a treatment system, for example, with a system or device for insulin therapy. Based on the target return path and / or risk metric defined for the registered glycemic state, computing device 66 is capable of adjusting, for example, basal rate and / or bolus as part of insulin therapy or other suitable treatment provided to a person. For example, insulin therapy can be adjusted so that the person’s return to the target glycemic state passes mainly along the target return path.

The risk metric values associated with the target return path defined for the registered glycemic state may be undesirable or may exceed the specified limits, and therefore treatment is adjusted so that the return to the target glycemic state proceeds in a different way. For example, it may be desirable to avoid the maximum penalty value identified with the target return path defined for the registered glycemic state due to the increased risk or danger to the person associated with this penalty value. For example, the maximum value of the fine may exceed a predetermined risk threshold established for a person. Moreover, the treatment can be adjusted so that during the person’s return to the target glycemic state, avoid getting into the glycemic state, where the maximum value of the penalty takes place. In this example, treatment can be adjusted with a focus on the second return path, following which eliminates getting into the glycemic state, which has a maximum penalty value.

Risk metrics for the history of glycemic indicators can be used retrospectively to analyze the behavior of a person with diabetes and draw conclusions based on them, as well as to determine the priorities on which to focus efforts on the treatment of diabetes. Such behavioral factors may include habits and diet, boluses, basal rates, physical activity, interventions to remove hypoglycemic or hyperglycemic conditions, corrective boluses, sleep indicators, etc. Risk metrics such as accumulated fines and average fines can be used to relate the behavioral factors of a person with diabetes to an increase in accumulated fines or average fines and thus identify behavioral factors that increase risk levels.

Although various options for devices, systems, methods and non-volatile computer-readable storage media for the analysis of glycemic state have been described in detail above, these options are given only as examples, not limiting all the possibilities of carrying out the invention. Therefore, it should be borne in mind that when implementing the invention, within the scope of its protection, various changes and improvements can be made to the examples considered, and elements of the objects of the invention can be replaced by their equivalents. The present description does not really exhaust the possibilities of carrying out the invention and does not limit the scope of its protection.

Further, the invention in certain embodiments can be presented as a method and / or process with a specific sequence of steps (steps). However, unless the steps described in the description are essential to the aforementioned method or process, the possibilities for implementing the method or process are not limited to the described sequence of steps. Other steps may be possible. Thus, the specific order of the steps of the method disclosed in the description should not be construed as limiting the scope of protection of the present invention. In addition, the disclosure related to the method and / or process should not be limited to the implementation of the appropriate steps in this order. Such sequence of steps may vary within the scope of protection of the present invention.

The following describes some embodiments of the invention. The technical solution proposed in the invention can also be implemented, for example, in a computer data processing system, a computing device or, in particular, in a device for measuring glycemia, having means for performing the steps of the methods described in the following embodiments of the invention.

1. The method of analysis of glycemic state, characterized in that:

- through at least one computing device determine the target glycemic state, which includes the target level of glycemia and the target rate of change of the target level of glycemia;

- by means of at least one computing device, the initial glycemic state is determined including the initial glycemic level and the initial rate of change of the initial glycemic level, the initial glycemic state being different from the target glycemic state;

- by means of the risk analysis logic available in at least one computing device, the target return path is calculated to transition from the initial glycemic state to the target glycemic state and containing at least one intermediate glycemic state associated with the transition from the initial glycemic state to the target glycemic state, wherein the target return path is calculated by the risk analysis logic based on, i.e. taking into account the risk associated with at least one intermediate glycemic condition on the target return path.

2. The method of embodiment 1, wherein the aforementioned calculation includes:

- determination between the initial glycemic state and the target glycemic state of a plurality of potential intermediate glycemic states; and

- the choice from the set of potential intermediate glycemic states of at least one intermediate glycemic state for the target return path from the condition of minimizing the risk associated with the target return path.

3. The method of embodiment 1 or 2, wherein the at least one intermediate glycemic state on the target return path includes a plurality of intermediate glycemic states, each of which has a corresponding penalty value, wherein the plurality of intermediate glycemic states are selected from the plurality of potential intermediate glycemic states from the condition of minimizing the sum of the fine values associated with the target return path, and each fine value includes a measure of risk associated with the corresponding intermediate glycemic state.

4. The method according to option 1, 2 or 3, in which the target return path is calculated by at least one computing device also taking into account the physiological limit of glycemic perturbation.

5. The method according to option 4, in which the physiological limit includes a given maximum acceleration of the level of glycemia.

6. The method of embodiment 4 or 5, wherein the aforementioned calculation includes: determining a plurality of potential glycemic states based on the target glycemic state, the physiological limit of glycemic perturbation, and a predetermined period for transition from each of the potential glycemic states to the target glycemic state; and selecting from a variety of potential glycemic states of an intermediate glycemic state for the target return path from the condition of minimizing the risk associated with the target return path.

7. The method according to any one of options 1-6, further characterized in that: by means of a risk analysis logic, a target return path for each of the plurality of initial glycemic states is calculated, each target return path containing a plurality of intermediate glycemic states associated with a transition from the initial glycemic state state to the target glycemic state and is calculated by at least one computing device from the condition of minimizing the risk associated with a multitude of intermediate glycemic states Nij this target return path; using the logic of risk analysis, a look-up table is formed that correlates each of the many initial glycemic conditions with the corresponding target return path; and storing a lookup table in a storage device accessible to at least one computing device.

8. The method according to option 7, also characterized in that: register, i.e. establish the glycemic state of a person suffering from diabetes based on at least one measured value of glycemia obtained using a glucose sensor; through the logic of risk analysis are addressed, i.e. accessing a look-up table to find among the plurality of initial glycemic states of that initial glycemic state, which basically coincides with the registered human glycemic state; and from the look-up table, the penalty value associated with the found initial glycemic state and representing the risk associated with the found initial glycemic state is determined taking into account the target return path corresponding to this initial glycemic state.

9. The method according to option 8, in which the registered glycemic state includes the level of glycemia, the rate of change of the level of glycemia and the uncertainty of at least one level of glycemia and its rate of change, and the method also includes calculating the risk associated with the registered glycemic state, based on the value of the fine and the specified uncertainty.

10. The method of embodiment 8 or 9, further comprising providing a fine value for display.

11. The method according to any of the previous options, in which the target glycemic state is the optimal glycemic state with a penalty value representing the risk associated with the target glycemic state and equal to zero.

12. The method according to any of the previous options, in which the target glycemic state includes a target glycemic level of approximately 112.5 milligrams per deciliter, and a target rate of change of about zero milligrams per deciliter per second.

13. A method for analyzing the glycemic state of a person suffering from diabetes, characterized in that: via at least one computing device, i.e. establish the glycemic state of a person on the basis of at least one measured value of glycemia obtained using a glucose sensor, and the recorded glycemic state includes a person’s glycemia level and the rate of change in glycemia level; using the risk analysis logic available in the at least one computing device, the return target path for transitioning from the registered glycemic state to the target glycemic state including the target glycemic level and the target rate of change of the target glycemic level is determined, the target return path containing at least at least one intermediate glycemic state associated with the transition from the registered glycemic state to the target glycemic state; and through the risk analysis logic available in the at least one computing device, at least one risk metric associated with the registered glycemic state is calculated based on at least one intermediate glycemic state on the target return path.

14. The method of embodiment 13, wherein the target return path contains a plurality of intermediate glycemic conditions, each of which has a corresponding penalty value, the registered glycemic state also has a corresponding penalty value, and each penalty value includes a risk measure associated with the corresponding glycemic state .

15. The method of embodiment 14, wherein the at least one risk metric associated with the registered glycemic state includes an accumulated penalty value comprising the sum of the penalty values associated with

with many intermediate glycemic conditions, and the value of the penalty associated with the registered glycemic state.

16. The method of embodiment 14, wherein the at least one risk metric associated with the registered glycemic state includes an average penalty ratio for the target return path calculated based on the relationship between the sum of the fine values associated with the target return path and the total the estimated time required to make the transition from the registered glycemic state to the target glycemic state along the target return path.

17. The method of embodiment 14, wherein the at least one risk metric associated with the registered glycemic state includes a maximum penalty of glycemic conditions on the target return path.

18. The method of embodiment 13, wherein the at least one risk metric associated with the registered glycemic state includes the total estimated time that the person transitions from the registered glycemic state to the target glycemic state along the target return path.

19. The method of embodiment 18, wherein the total estimated time is calculated based on a predetermined maximum glycemic acceleration and the number of intermediate glycemic states on the target return path.

20. The method of embodiment 13, further comprising correcting the treatment of the person through at least one computing device to transfer the person from the registered glycemic state to the target glycemic state based on the target return path.

21. The method according to option 20, in which the correction of treatment includes the correction of the basal rate of supply of insulin and / or bolus in case of insulin therapy prescribed to a person.

22. The method according to embodiment 13, also characterized in that: by means of the risk analysis logic, at least one risk metric associated with the registered glycemic state is compared with a risk threshold; and if the risk threshold exceeds at least one risk metric, the person’s treatment is adjusted to transfer the person from the registered glycemic state to the target glycemic state according to a second return path different from the target return path by at least one intermediate glycemic state.

23. The method of embodiment 13, further characterized in that: a plurality of human glycemic states are recorded for determining a history of human glycemic states, each glycemic state including a person’s glycemic level and a rate of change in glycemic level; for each registered glycemic state, a target return path from the registered glycemic state to the target glycemic state is determined, wherein each target return path contains at least one intermediate glycemic state associated with a transition from the registered glycemic state to the target glycemic state; and for each registered glycemic state, at least one risk metric associated with the registered glycemic state is calculated based on at least one intermediate glycemic state on the target return path.

24. The method of embodiment 13, wherein the target glycemic state includes a target glycemic level of about 112.5 milligrams per deciliter, and a target rate of change of about 0 milligrams per deciliter per second.

25. The method of embodiment 13, wherein the registered glycemia level and / or its registered rate of change is estimated by at least one computing device based on at least one measured glycemia value, weighted by the probability of accuracy of the glucose sensor readings.

26. The method according to option 25, in which the registered level of glycemia and / or the registered rate of change is estimated using a recursive filter available in at least one computing device.

27. The method of embodiment 25, wherein the at least one risk metric associated with the registered glycemic state is calculated based on the penalty value associated with the registered glycemic state and based on the accuracy of the glucose sensor.

28. The method of embodiment 13, wherein each of the following parameters is recorded when registering a person’s glycemic state: the level of glycemia, the rate of change in the level of glycemia, and the acceleration of the level of glycemia, based on at least one measured value of glycemia obtained using a glucose sensor.

29. The method of embodiment 13, which also uses a look-up table correlating each of the many glycemic states with a corresponding target return path and a corresponding risk metric, the look-up table being stored in a storage device accessible to at least one computing device, and computing at least one risk metric associated with a registered glycemic state includes a lookup table to find a glycemic s a condition essentially coinciding with the registered glycemic state of a person, and extracting from the look-up table a risk metric corresponding to the found glycemic state.

30. A non-volatile computer-readable storage medium containing executable instructions, the execution of which by at least one processor ensures that said at least one processor performs the following actions: determining a target glycemic state including a target glycemic level and a target rate of change of the target glycemic level; determination of the initial glycemic state, which includes the initial level of glycemia and the initial rate of change of the initial level of glycemia, and the initial glycemic state differs from the target glycemic state; calculating a target return path intended to transition from the initial glycemic state to the target glycemic state and containing at least one intermediate glycemic state associated with the transition from the initial glycemic state to the target glycemic state, the target return path being calculated by at least one processor based on , i.e. taking into account the risk associated with at least one intermediate glycemic condition on the target return path.

31. The non-volatile computer-readable storage medium according to embodiment 30, wherein, when said instructions are executed, at least one processor also performs the following actions: determining between the initial glycemic state and the target glycemic state of a plurality of potential intermediate glycemic states; and selecting from the plurality of potential intermediate glycemic states of at least one intermediate glycemic state for the target return path from the condition of minimizing the risk associated with the target return path.

32. The non-volatile computer-readable storage medium of embodiment 31, wherein the at least one intermediate glycemic state on the target return path includes a plurality of intermediate glycemic states, each of which has a corresponding penalty value, wherein the plurality of intermediate glycemic states are selected by at least one processor from the set of potential intermediate glycemic states from the condition of minimizing the sum of the fine values associated with the target return path fee, and each penalty value includes a measure of risk associated with the corresponding intermediate glycemic state.

33. Non-volatile computer-readable storage medium according to option 30, in which the target return path is calculated by at least one processor also taking into account the physiological limit of glycemic perturbation.

34. Non-volatile computer-readable storage medium according to option 33, in which when these instructions are executed, at least one processor also performs the following actions: determining a set of potential glycemic states based on the target glycemic state, physiological limit of glycemic perturbation and a given period for transition to the target glycemic state from each of the potential glycemic conditions; and selecting from a variety of potential glycemic states of an intermediate glycemic state for the target return path from the condition of minimizing the risk associated with the target return path.

35. The non-volatile computer-readable storage medium according to embodiment 30, wherein, when said instructions are executed, at least one processor also performs the following actions: calculating a target return path for each of the plurality of initial glycemic states, each target return path containing a plurality of intermediate glycemic states associated with the transition from the initial glycemic state to the target glycemic state, and is calculated from the condition of minimizing the risk associated with a multitude of intermediate full-time glycemic state on the target path of the return; generating a look-up table correlating each of the plurality of initial glycemic states with a corresponding target return path, and storing the look-up table in a storage device accessible to at least one computing device.

36. Non-volatile computer-readable storage medium according to option 35, in which when these instructions are executed, at least one processor also performs the following actions: registration of the glycemic state of a diabetic person based on at least one measured glycemia value obtained using a glucose sensor; referring to the look-up table to find among the set of initial glycemic states of that initial glycemic state, which basically coincides with the registered glycemic state of a person; and determining from the look-up table the value of the penalty associated with the found initial glycemic state and representing the risk associated with the found initial glycemic state, taking into account the target return path corresponding to that initial glycemic state.

37. Non-volatile computer-readable storage medium containing executable instructions, the execution of which by at least one processor ensures that the specified at least one processor performs the following actions: registration of a person’s glycemic state based on at least one measured glycemia value obtained using a glucose sensor moreover, the registered glycemic state includes the level of human glycemia and the rate of change in the level of glycemia; determining a target return path for transitioning from a registered glycemic state to a target glycemic state including a target glycemic level and a target rate of change of a target glycemic level, the target return path containing at least one intermediate glycemic state associated with a transition from a registered glycemic state to target glycemic state; and calculating at least one risk metric associated with the registered glycemic state based on at least one intermediate glycemic state on the target return path.

38. The non-volatile computer-readable storage medium of embodiment 37, wherein the target return path contains a plurality of intermediate glycemic states, each of which has a corresponding penalty value, the registered glycemic state also has a corresponding penalty value, and each penalty value includes a risk measure associated with corresponding glycemic state.

39. The non-volatile machine-readable storage medium of embodiment 38, wherein the at least one risk metric associated with the registered glycemic state includes at least one of the following: accumulated penalty value for the target return path, average penalty coefficient for the target return path and the maximum penalty for glycemic states on the target return path, and the accumulated penalty contains the sum of the fine values associated with the set of intermediate glycemic states, and penalties associated with the registered glycemic state, and the average penalty ratio is based on the ratio between the sum of the fine values and the total estimated time required to complete the transition from the registered glycemic state to the target glycemic state along the target return path.

40. The non-volatile computer-readable storage medium according to embodiment 37, wherein the at least one risk metric associated with the registered glycemic state includes the total estimated time of transition of the person from the registered glycemic state to the target glycemic state along the target return path.

41. The non-volatile computer-readable storage medium according to embodiment 37, wherein, when said commands are executed, at least one processor also adjusts the person’s treatment to transfer the person from the registered glycemic state to the target glycemic state based on the target return path, and the treatment correction includes basal correction insulin and / or bolus delivery rates for insulin therapy prescribed to humans.

42. The non-volatile computer-readable storage medium according to embodiment 37, wherein, when said instructions are executed, at least one processor also estimates the glycemia level and / or rate of change of the glycemia level based on at least one measured glycemia value weighted by the accuracy of the glucose sensor.

43. The non-volatile computer-readable storage medium of embodiment 42, wherein, when said instructions are executed, at least one processor also calculates the risk associated with the registered glycemic state based on the penalty value associated with the registered glycemic state and based on the accuracy of the glucose sensor.

44. Non-volatile machine-readable storage medium according to option 37, in which to calculate at least one risk metric associated with a registered glycemic state, when performing these commands, at least one processor accesses a look-up table correlating each of the many glycemic states with the corresponding target by returning and with the appropriate risk metric to find in the look-up table that glycemic state that essentially coincides with the registered gly emicheskim human condition, and extracts from the look-up table of risk metrics corresponding to the found glycemic status.

Claims (55)

1. A computing device for a continuous glycemic monitoring system comprising at least one processor, a storage device accessible to the processor, and an input device electrically coupled to the processor, the processor comprising risk analysis logic configured to implement a method for analyzing glycemic state, characterized in , what using a computing device determines the target glycemic state, which includes the target level of glycemia and the target rate of change of the target level of glycemia; by means of a computing device, the initial glycemic state is determined, which includes the initial glycemic level and the initial rate of change of the initial glycemic level, the initial glycemic state being different from the target glycemic state; using the risk analysis logic available in the computing device, a return path is designed to transition from the initial glycemic state to the target glycemic state and containing at least one intermediate glycemic state associated with the transition from the initial glycemic state to the target glycemic state, the path return is calculated through risk analysis logic based on the risk of developing hypo- or hyperglycemia associated with at least one them an intermediate glycemic state on the path of return; using a computing device, the person’s treatment is adjusted to transfer the person from the registered glycemic state to the target glycemic state based on the return path and / or risk metric for the development of hypo- or hyperglycemia, moreover, at least one risk metric associated with the registered glycemic state includes the total estimated time of the person’s transition from the registered glycemic state to the target glycemic state along the return path, and the total estimated time is calculated based on the specified maximum glycemic acceleration and the number of intermediate glycemic states on the way to return. 2. The computing device according to claim 1, in which the aforementioned calculation includes: determining between the initial glycemic state and the target glycemic state of a plurality of potential intermediate glycemic states; and selection from among many potential intermediate glycemic conditions of at least one intermediate glycemic state for the return path from the condition of minimizing the risk of developing hypo- or hyperglycemia associated with the return path. 3. The computing device of claim 1 or 2, wherein the at least one intermediate glycemic state on the return path includes a plurality of intermediate glycemic states, each of which has a corresponding penalty value, wherein the plurality of intermediate glycemic states are selected from a plurality of potential intermediate glycemic conditions from the condition of minimizing the sum of the fine values associated with the return path, and each fine value includes a measure of the risk of developing hypo- or hyperglycemia, associated with the corresponding intermediate glycemic state. 4. The computing device according to claim 1 or 2, further configured to calculate the return path also taking into account a given maximum acceleration of the glycemia level. 5. The computing device according to claim 4, in which the aforementioned calculation includes: determination of the set of potential glycemic states on the basis of the target glycemic state, a given maximum acceleration of the glycemic level and a given period for the transition from each of the potential glycemic states to the target glycemic state; and the choice of a set of potential glycemic states of an intermediate glycemic state for the return path from the condition of minimizing the risk of developing hypo- or hyperglycemia associated with the return path. 6. The computing device according to one of paragraphs. 1, 2, 4, 5, also configured to calculating, by means of risk analysis logic, the return path for each of the many initial glycemic states, each return path containing many intermediate glycemic states associated with the transition from the initial glycemic state to the target glycemic state, and is calculated by a computing device from the condition of minimizing the risk of developing hypo- or hyperglycemia associated with many intermediate glycemic conditions on this return path; forming, through the logic of risk analysis, a look-up table correlating each of the many initial glycemic conditions with the corresponding return path; and storing the lookup table in a storage device accessible to the computing device. 7. The computing device according to claim 6, also made with the possibility registration of the glycemic state of a person suffering from diabetes based on at least one measured value of glycemia obtained using a glucose sensor that provides continuous monitoring of glycemia; accessing the risk analysis logic to a look-up table to find among the set of initial glycemic states that initial glycemic state, which basically coincides with the registered human glycemic state; and determining, according to the look-up table, the value of the penalty associated with the found initial glycemic state and representing the risk of developing hypo- or hyperglycemia associated with the found initial glycemic state, taking into account the return path corresponding to this initial glycemic state. 8. The computing device according to claim 7, in which the registered glycemic state includes the level of glycemia, the rate of change of the level of glycemia and the uncertainty of at least one level of glycemia and its rate of change, and the computing device is configured to calculate the risk of developing hypo- or hyperglycemia associated with the registered glycemic state, based on the value of the fine and the specified uncertainty. 9. The computing device according to claim 7 or 8, also configured to provide a fine value for display on a display and displaying a fine value on a display accessible to a computing device. 10. The computing device according to one of paragraphs. 1, 2, 4, 5, 7, 8, in which the target glycemic state is the optimal glycemic state with a penalty value representing the risk of developing hypo- or hyperglycemia associated with the target glycemic state and equal to zero. 11. The computing device according to one of paragraphs. 1, 2, 4, 5, 7, 8, in which the target glycemic state includes a target glycemic level of about 112.5 milligrams per deciliter, and a target rate of change of about 0 milligrams per deciliter per second. 12. The method of analysis of the glycemic state of a person suffering from diabetes, characterized in that a glycemic state of a person is recorded by a computing device of a continuous glycemic monitoring system based on at least one measured glycemic value obtained with a glucose sensor that provides continuous glycemic monitoring, wherein the recorded glycemic state includes a person’s glycemic level and rate of change in glycemic level; by means of the risk analysis logic available in the computing device, a return path is determined for moving from the registered glycemic state to a target glycemic state including a target glycemic level and a target rate of change of the target glycemic level, the return path containing at least one intermediate glycemic state, associated with the transition from the registered glycemic state to the target glycemic state; and using the risk analysis logic available in the computing device, at least one risk metric for the development of hypo- or hyperglycemia associated with the registered glycemic state is calculated based on at least one intermediate glycemic state on the return path; using a computing device, the person’s treatment is adjusted to transfer the person from the registered glycemic state to the target glycemic state based on the return path and / or risk metric for the development of hypo- or hyperglycemia, moreover, at least one risk metric associated with the registered glycemic state includes the total estimated time of the person’s transition from the registered glycemic state to the target glycemic state along the return path, and the total estimated time is calculated based on the specified maximum glycemic acceleration and the number of intermediate glycemic states on the way to return. 13. The method according to p. 12, in which the return path contains many intermediate glycemic conditions, each of which has a corresponding penalty value, the registered glycemic state also has a corresponding penalty value, and each penalty value includes a risk measure for developing hypo- or hyperglycemia, associated with the corresponding glycemic state. 14. The method of claim 13, wherein the at least one risk metric associated with the registered glycemic state includes a cumulative penalty value comprising the sum of the penalty values associated with a plurality of intermediate glycemic conditions and the penalty value associated with the registered glycemic state condition. 15. The method of claim 13 or 14, wherein the at least one risk metric associated with the registered glycemic state includes an average penalty rate for a return path calculated based on the relationship between the sum of the penalty values associated with the return path, and The total estimated time required to complete the transition from the registered glycemic state to the target glycemic state along the return path. 16. The method of claim 13 or 14, wherein the at least one risk metric associated with the registered glycemic state includes a maximum penalty for glycemic conditions on the return path. 17. The method according to one of paragraphs. 1, 12-14, in which the correction of a person’s treatment performed by means of a computing device for transferring a person from a registered glycemic state to a target glycemic state is based on the return path. 18. The method according to p. 17, in which the correction of treatment includes the correction of the basal rate of supply of insulin and / or bolus for insulin therapy prescribed to a person. 19. The method according to one of paragraphs. 1, 12-14, 18, also characterized in that using the risk analysis logic, at least one risk metric associated with the registered glycemic state is compared with a risk threshold; and if the risk threshold exceeds at least one risk metric, the person’s treatment is adjusted to transfer the person from the registered glycemic state to the target glycemic state according to the second return path, which differs from the return path by at least one intermediate glycemic state. 20. The method according to one of paragraphs. 1, 12-14, 18, also characterized in that registering a lot of human glycemic conditions to determine the history of human glycemic conditions, containing a series of recorded glycemic levels measured over a period of time, each glycemic state includes a person’s glycemic level and rate of change in glycemic level; for each registered glycemic state, a return path from the registered glycemic state to the target glycemic state is determined, each return path containing at least one intermediate glycemic state associated with the transition from the registered glycemic state to the target glycemic state; and for each registered glycemic state, at least one risk metric associated with the registered glycemic state is calculated based on at least one intermediate glycemic state on the return path. 21. The method according to one of paragraphs. 1, 12-14, 18, in which the target glycemic state includes a target glycemic level of about 112.5 milligrams per deciliter, and a target rate of change of about 0 milligrams per deciliter per second. 22. The method according to one of paragraphs. 1, 12-14, 18, in which the registered level of glycemia and / or the registered rate of change is estimated by at least one computing device based on at least one measured value of glycemia, weighted by the probability of accuracy of the readings of the glucose sensor, providing continuous monitoring glycemia. 23. The method according to p. 22, in which the registered level of glycemia and / or the registered rate of change is estimated using a recursive filter available in at least one computing device. 24. The method of claim 23, wherein the at least one risk metric associated with the registered glycemic state is calculated based on the penalty value associated with the registered glycemic state and based on the accuracy of the glucose sensor providing continuous glycemic monitoring. 25. The method according to one of paragraphs. 1, 12-14, 18, 23, 24, in which when registering a person’s glycemic state, each of the following parameters is recorded: the level of glycemia, the rate of change in the level of glycemia and the acceleration of the level of glycemia, based on the measured values of glycemia obtained using the glucose sensor, providing continuous monitoring of glycemia. 26. The method according to one of paragraphs. 1, 12-14, 18, 23, 24, which also uses a look-up table correlating each of the many glycemic states with a corresponding return path and a corresponding risk metric, the look-up table being stored in a storage device accessible for at least one computing devices, and the calculation of at least one risk metric associated with a registered glycemic state includes: referring to the look-up table to find the glycemic state corresponding to the registered glycemic state of a person in the look-up table, and extracting from the look-up table a risk metric corresponding to the found glycemic state. 27. A computer-readable storage medium containing executable instructions, when executed by at least one processor, said at least one processor performs the method according to one of claims. 12-26. 28. A computer system containing means for performing the steps of the methods according to one of paragraphs. 12-26.

Images